hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Vladimir Rodionov (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-15513) hbase.hregion.memstore.chunkpool.maxsize is 0.0 by default
Date Fri, 02 Sep 2016 14:51:20 GMT

    [ https://issues.apache.org/jira/browse/HBASE-15513?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15458734#comment-15458734

Vladimir Rodionov commented on HBASE-15513:

Ya at least the tests what we did using CMS and G1 shows that chunk pool still have value.
With G1GC and with out chunk pool, we were able to get latency and GC pause same as with pool
case BUT with much more heap size allocated for the RS. So said that, the chunk pool was helping
us to achieve what we wanted with lower heap size. Also to mention that here we set global
memstore upper limit as 43% only as we wanted to keep the IHOP for G1GC as some thing like
50% max.

I have run my own as well and measured memory allocation rate (not latency) and major GC frequency.

Good improvements in Old and Young GC times. It reduced Old GC events by 50% and time by 60%.

> hbase.hregion.memstore.chunkpool.maxsize is 0.0 by default
> ----------------------------------------------------------
>                 Key: HBASE-15513
>                 URL: https://issues.apache.org/jira/browse/HBASE-15513
>             Project: HBase
>          Issue Type: Sub-task
>    Affects Versions: 2.0.0
>            Reporter: Vladimir Rodionov
>            Assignee: Vladimir Rodionov
>             Fix For: 2.0.0
>         Attachments: HBASE-15513-v1.patch
> That results in excessive MemStoreLAB chunk allocations because we can not reuse them.
Not sure, why it has been disabled, by default. May be the code has not been tested well?

This message was sent by Atlassian JIRA

View raw message